Steps to correct the problem (some of these steps may, or may not, have
helped):
1. Update the linux kernel-header files.
2. Order MATLAB 2007a+ (and install, with some incredible help from
MATLAB's Super-Duper Customer Service)
3. Redownload spm5.tar.gz and Updates_826.tar.gz.
4. Reinstall spm5 and update package.
5. Recompile the mexa64 files as per Kathy Pearson's wondeful
suggestions. (Contained in this thread, I think she attached a file
spm5_mex.txt)
6. Install compat-libstdc++ for the x86_64 distro.
Off to the races!! Yay!
I think that the redownload of spm5 (and updates) was the crucial step,
although, I tried so many things that it's hard to say :)
THANKYOU ALL FOR ALL YOUR HELP! You've been invaluable; once again, I'm
impressed by the community behind SPM (and all the other free imaging
softwares out there).
Talk to you all soon, when I have my next intractable problem....
-Daniel
>>> John Ashburner <[log in to unmask]> 07/30/07 10:54 am >>>
This is very strange. If you do Control-C, does MATLAB print out a
stack of
things that it is trying to do? This snippet may help to narrow down
the
problem...
spm_defaults;
spmdir = fileparts(which('spm'));
fname = fullfile(spmdir, 'templates', 'EPI.nii');
N=nifti(fname);
N.hdr
N
N.mat
All the best,
-John
On Wednesday 25 July 2007 00:02, Daniel Schwartz wrote:
> Thanks again, sorry to keep coming back at you...
>
> #1 hangs forever.
> #2 The process you've described (copying line by line into the Command
> WindowThe equivalent of using F10 to step through the spm_vol_nifti
> function, I assume, hangs on line 28, mat = N.mat. In fact, any
> attempt to access this particular field of object N hangs forever,
> including size, class, etc. However, the object doesn't seem unstable,
> as I can access other fields, such as N.dat, without trouble.
>
> I feel like there is something simple I am missing...
>
> Thankyou!
> -Daniel
>
>
>
> Daniel Schwartz, CIC
> Functional Brain Imaging Lab
> Oregon Health & Sciences University
> 3181 SW Sam Jackson Pk. Rd.
> Portland, OR 97239
> (503) 494-6494
>
> >>> Kathy Pearson <[log in to unmask]> 07/24/07 10:51 AM >>>
>
> Daniel,
>
> Here are some lines you might try:
> (1)
> matlab
> spm_defaults;
> spmdir = fileparts(which('spm'));
> fname = fullfile(spmdir, 'templates', 'EPI.nii');
> v = spm_vol(fname);
> vol = spm_read_vols(v);
>
> If #1 fails, then if you think the problem is in spm_vol_nifti:
> (2)
> matlab
> spm_defaults;
> spmdir = fileparts(which('spm'));
> fname = fullfile(spmdir, 'templates', 'EPI.nii');
> n = [1 1];
>
> ... drag and drop every matlab statement AFTER this one
> if nargin<2, n = [1 1]; end;
> ... from spm_vol_nifti to your matlab command window
>
> That might give you some info. Also, you may want to use some other
> file name from the templates directory or elsewhere.
>
> If you have trouble when you reference this line in spm_vol_nifti:
> N = nifti(fname);
> that may indicate that the SPM version 5 code is not on top in your
> matlab path. If your SPM5 path is here:
> /code/spm5
> and other SPM versions exist, then these commands may be needed for
> setup:
> addpath('/code/spm5');
> spmver = spm('ver', [], 1);
> disp(['setting paths for version ... ' spmver]);
> spm_defaults;
>
> Kathy Pearson
> WFUHS Radiology
>
> -----Original Message-----
> From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]
> On Behalf Of Daniel Schwartz
> Sent: Tuesday, July 24, 2007 1:07 PM
> To: [log in to unmask]
> Subject: Re: [SPM] Still can't get SPM5 to Display... Help?
>
> Thankyou all for your help.
>
> I still can't get SPM5 to display a single file. It still hangs,
exactly
> like it did before. However, I followed Ged Ridgeway's advice (after
> recompiling the mex functions as per Kathy Pearson's advice) and tried
> to step through each function. It seems that it hangs on every call to
> access the array N.mat. In particular, when I try to load EPI.nii from
> the templates folder, it hangs on line 28 of spm_vol_nifti, mat =
> N.mat. Calls to most of the other fields in the N object seem to go
> just fine, such as X = double(N.dat); which is a fairly large array,
> (~90x90x90), so I don't think that the cpu is just getting drowned by
> the sheer array sizes or anything. Does anyone have any advice about
> getting SPM5 working on my poor unfortunate Linux box?
>
> Commands to reproduce my error:
> > file = fullfile(spm('dir'), 'templates', 'EPI.nii');
> > dbstop in spm_vol_nifti
> > vol = spm_vol(file)
> > <<F10 to line 28>>
>
> Fedora Core 6 (Zod), Dual Quad Core Intel 2.33 GHz Processors, 4 GB
RAM,
> gcc v4.1.2, MATLAB R2006b.
>
> Thankyou! My modified Makefile is attached...
> -Daniel
>
>
>
>
>
>
>
> Daniel Schwartz, CIC
> Functional Brain Imaging Lab
> Oregon Health & Sciences University
> 3181 SW Sam Jackson Pk. Rd.
> Portland, OR 97239
> (503) 494-6494
>
> >>> Kathy Pearson <[log in to unmask]> 07/11/07 8:04 AM >>>
>
> Daniel,
>
> Attached are the commands that I use to recompile the MEX functions
for
> RedHat 64-bit Linux. Note that your matlab path may be different.
>
> Kathy Pearson
> WFUHS Radiology
>
> -----Original Message-----
> From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]]
> On Behalf Of Daniel Schwartz
> Sent: Tuesday, July 10, 2007 3:04 PM
> To: [log in to unmask]
> Subject: [SPM] Still can't get SPM5 to Display... Help?
>
> MATLAB R2006b 64-bit, SPM5, Fedora Core 6 (linux) x86-64, Intel Dual
> Quad-core 2.33 Ghz, 4GB RAM.
>
> Hello all!
>
> I've downloaded latest updates, updated (overwritten) all the
applicable
> files, and run make && make install from ~/spm5/src. I start MATLAB, I
> run spm fmri, I click Display, I choose ~/spm5/templates/EPI.nii, And
> I just get a spinner forever. Does anyone have any advice? There is no
> error message printed at all and I don't know mex, otherwise I'd chase
> this down myself.
>
> Thankyou for any and all help...
> -Daniel Schwartz
